Motion estimation method by employing a stochastic sampling technique
View Patent ↗In a motion estimation method for use in encoding a moving picture, a full-pixel motion vector is estimated by stochastically sampling a pixel to be processed in a predetermined-sized block of a previous frame or a next frame as a reference frame for each of a plurality of equal-sized blocks in a current frame. Then, a half-pixel motion vector is estimated based on the full-pixel motion vector. Accordingly, both the calculation amount and the calculation time required for the motion estimation are effectively reduced. Further, it can be prevented that the hardware becomes complicated.
1. A motion estimation method for use in encoding a moving picture, comprising:
(a) estimating a full-pixel motion vector by stochastically sampling pixels to be processed in a predetermined-sized block of a reference frame for each of a plurality of equal-sized blocks in a current frame; and
(b) estimating a half-pixel motion vector based on the full-pixel motion vector,
wherein the pixels in the predetermined-sized block in the step (a) are sampled by using the following equation:
R ( t+ 1)= R ( t )+ r+c+ 1, C ( t+ 1)= C ( t )+ r+c+ 1,
wherein r represents a probability variable for use in sampling a pixel on the X-axis,
wherein the probabilities of r being “1” and “0” are respectively ½;
wherein c represents a probability variable for use in sampling a pixel on the Y-axis;
wherein the probabilities of c being “1” and “0” are respectively ½;
wherein R(t) and C(t) represents a rotation coefficient of a currently sampled pixel on the X-axis and on the Y-axis. respectively; and
wherein each of R(t+1) and C(t+1) represents a location coefficient of a pixel to be sampled next.
2. The method of claim 1 , wherein the pixels in the predetermined sized block are sampled in the step (a) with the probability of 1/n (n being a positive integer larger than 1) for both an X-axis and an Y-axis.
3. The method of claim 2 , wherein the pixels in the predetermined-sized block are sampled in the step (a) with the probability of ½ for both the X-axis and the Y-axis.
4. The method of claim 1 , wherein a sum of absolute difference of a pixel s(xy)(t−1) in a previous frame corresponding to a pixel s(xy) (t) of a current coordinate (x,y) is obtained by the following equation:
∑
x
N
∑
y
N
[
s
(
x
,
y
)
(
t
)
-
s
(
x
,
y
)
(
t
-
1
)
·
I
(
x
,
y
)
]
=
1
4
∑
x
N
∑
y
N
[
s
(
x
,
y
)
(
t
)
-
s
(
x
,
y
)
(
t
-
1
)
]
+
ɛ
(
t
)
wherein N represents a number of pixels on the X-axis or on the Y-axis in a block and I(x,y) represents a characterizing function for showing whether the pixel in the current coordinate (x,y) is sampled or not.
5. A motion estimation method for use in encoding a moving picture, comprising:
(a) estimating a full-pixel motion vector by stochastically sampling pixels to be processed in a predetermined-sized block of a reference frame for each of a plurality of equal-sized blocks in a current frame; and
(b) estimating a half-pixel motion vector based on the full-pixel motion vector,
wherein the pixels in the predetermined-sized block are sampled in the step (a) with the probability of ½ for both the X-axis and the Y-axis,
wherein the pixels in the predetermined-sized block in the step (a) are sampled by using the following equation:
R ( t+ 1)= R ( t )+ r+c+ 1, C ( t+ 1)= C ( t )+ r+c+ 1,
wherein r represents a probability variable for use in sampling a pixel on the X-axis,
wherein the probabilities of r being “1” and “0” are respectively ½;
wherein c represents a probability variable for use in sampling a pixel on the Y-axis;
wherein the probabilities of c being “1” and “0” are respectively ½;
wherein R(t) and C(t) represents a rotation coefficient of a currently sampled pixel on the X-axis and on the Y-axis, respectively; and
wherein each of R(t+1) and C(t+1) represents a location coefficient of a pixel to be sampled next.
6. A motion estimation method for use in encoding a moving picture, comprising:
estimating a full-pixel motion vector by stochastically sampling pixels to be processed in a predetermined-sized block of a reference frame for each of a plurality of equal-sized blocks in a current frame; and
estimating a half-pixel motion vector based on the full-pixel motion vector,
wherein a sum of absolute difference of a pixel s(xy)(t−1) in a previous frame corresponding to a pixel s(xy)(t) of a current coordinate (x,y) is obtained by the following equation:
∑
x
N
∑
y
N
[
s
(
x
,
y
)
(
t
)
-
s
(
x
,
y
)
(
t
-
1
)
·
I
(
x
,
y
)
]
=
1
4
∑
x
N
∑
y
N
[
s
(
x
,
y
)
(
t
)
-
s
(
x
,
y
)
(
t
-
1
)
]
+
ɛ
(
t
)
wherein N represents a number of pixels on the X-axis or on the Y-axis in a block and I(x,y) represents a characterizing function for showing whether the pixel in the current coordinate (x,y) is sampled or not.